Central and South Italian K15 results

[Blanco]

From Tuscany, Central Italy (HGDP sample) Kit Number: Z481041

#	Population	Percent
1	Atlantic	26.49
2	East_Med	24.34
3	West_Med	22.66
4	North_Sea	15.58
5	West_Asian	6.88
6	Baltic	2.34
7	Red_Sea	1.71

Single Population Sharing:

#	Population (source)	Distance
1	Tuscan	9.02
2	North_Italian	9.47
3	West_Sicilian	11.6
4	Spanish_Andalucia	12.77
5	Spanish_Murcia	13.04
6	Italian_Abruzzo	13.57
7	Spanish_Extremadura	13.58
8	Spanish_Valencia	13.88
9	Portuguese	14.5
10	Spanish_Castilla_La_Mancha	14.88
11	Spanish_Cataluna	15.12
12	Spanish_Castilla_Y_Leon	15.21
13	Spanish_Aragon	16.05
14	South_Italian	16.18
15	East_Sicilian	16.24
16	Spanish_Galicia	16.25
17	Greek	16.57
18	Central_Greek	16.62
19	Greek_Thessaly	17.28
20	Spanish_Cantabria	17.4

Mixed Mode Population Sharing:

#		Primary Population (source)	Secondary Population (source)	Distance
1		53.5%	Spanish_Aragon	+	46.5%	Italian_Jewish	@	6.18
2		58.9%	Italian_Jewish	+	41.1%	French_Basque	@	6.18
3		70.5%	Spanish_Aragon	+	29.5%	Lebanese_Druze	@	6.19
4		57.9%	Spanish_Valencia	+	42.1%	Italian_Jewish	@	6.27
5		55.1%	Spanish_Aragon	+	44.9%	Algerian_Jewish	@	6.46
6		59.5%	Spanish_Valencia	+	40.5%	Algerian_Jewish	@	6.53
7		76.3%	Spanish_Andalucia	+	23.7%	Lebanese_Druze	@	6.53
8		70.4%	Spanish_Aragon	+	29.6%	Lebanese_Christian	@	6.55
9		65%	Spanish_Aragon	+	35%	Cyprian	@	6.58
10		74.3%	Spanish_Valencia	+	25.7%	Lebanese_Druze	@	6.58
11		60.7%	Spanish_Andalucia	+	39.3%	Italian_Jewish	@	6.6
12		50.2%	Spanish_Aragon	+	49.8%	South_Italian	@	6.67
13		74.1%	Spanish_Valencia	+	25.9%	Lebanese_Christian	@	6.69
14		56%	Spanish_Castilla_La_Mancha	+	44%	Italian_Jewish	@	6.75
15		69.1%	Spanish_Valencia	+	30.9%	Cyprian	@	6.77
16		57.6%	Algerian_Jewish	+	42.4%	French_Basque	@	6.88
17		72.8%	Spanish_Castilla_La_Mancha	+	27.2%	Lebanese_Druze	@	6.88
18		54.9%	Spanish_Valencia	+	45.1%	South_Italian	@	6.89
19		62.4%	Spanish_Andalucia	+	37.6%	Algerian_Jewish	@	6.91
20		60.9%	West_Sicilian	+	39.1%	Spanish_Aragon	@	6.92

From Catanzaro, South Italy: Kit Number: A552689

#	Population	Percent
1	East_Med	33.78
2	Atlantic	15.33
3	West_Med	15.19
4	West_Asian	15.1
5	Baltic	7.49
6	Red_Sea	6.86
7	North_Sea	5.09
8	Southeast_Asian	0.89
9	Northeast_African	0.25
10	Eastern_Euro	0.01

Single Population Sharing:

#	Population (source)	Distance
1	South_Italian	5.98
2	Sephardic_Jewish	6.01
3	Italian_Jewish	7.31
4	Algerian_Jewish	7.9
5	East_Sicilian	8.01
6	Central_Greek	8.37
7	Tunisian_Jewish	9.08
8	West_Sicilian	9.81
9	Ashkenazi	9.86
10	Italian_Abruzzo	10.15
11	Cyprian	10.99
12	Greek	12.54
13	Libyan_Jewish	13.38
14	Lebanese_Muslim	14.45
15	Greek_Thessaly	14.66
16	Tuscan	15.69
17	Syrian	15.89
18	Samaritan	17.24
19	Turkish	17.88
20	Lebanese_Christian	18.53

Mixed Mode Population Sharing:

#		Primary Population (source)	Secondary Population (source)	Distance
1		83.5%	South_Italian	+	16.5%	Lebanese_Druze	@	4.66
2		89.8%	South_Italian	+	10.2%	Yemenite_Jewish	@	4.68
3		67.9%	South_Italian	+	32.1%	Tunisian_Jewish	@	4.72
4		83.4%	South_Italian	+	16.6%	Palestinian	@	4.74
5		50.4%	South_Italian	+	49.6%	Sephardic_Jewish	@	4.77
6		74.8%	Cyprian	+	25.2%	Spanish_Valencia	@	4.8
7		83.6%	South_Italian	+	16.4%	Lebanese_Christian	@	4.85
8		80.7%	Cyprian	+	19.3%	French_Basque	@	4.85
9		82.7%	South_Italian	+	17.3%	Samaritan	@	4.87
10		77.3%	Cyprian	+	22.7%	Southwest_French	@	4.9
11		53.6%	West_Sicilian	+	46.4%	Cyprian	@	4.91
12		77.1%	Sephardic_Jewish	+	22.9%	Greek	@	4.93
13		75.1%	Cyprian	+	24.9%	Spanish_Castilla_Y_Leon	@	4.94
14		76.5%	Cyprian	+	23.5%	Spanish_Aragon	@	4.95
15		80.2%	South_Italian	+	19.8%	Lebanese_Muslim	@	4.96
16		74.8%	South_Italian	+	25.2%	Cyprian	@	4.99
17		74%	Cyprian	+	26%	Spanish_Andalucia	@	5.03
18		91.1%	South_Italian	+	8.9%	Saudi	@	5.06
19		67.3%	Sephardic_Jewish	+	32.7%	Central_Greek	@	5.07
20		85%	South_Italian	+	15%	Jordanian	@	5.07

 

[Angela]

Well, goodness, if that really is a Tuscan sample from HGDP, then this really is a really TERRIBLE calculator. A genetic distance of over 9? When you're dealing with moderns, if the distance isn't at least 4 or 5 forget it.

This must be as bad as Polako's "J" calculator where everybody has huge percentages of "Jewish" ancestry.

Like I said, some amateurs, especially those from eastern Europe, have Jews on the brain.

 

[davef]

Right, Angela, and it's weird that Ashkenazi are scoring 26 percent ashkenazi on that, there's some weird sorcery behind that thing we may never figure out. Normal calculators with an ashkenazi component give 100 percent or perhaps a bit less of that to ashkenazi users.

 

[Angela]

Right, Angela, and it's weird that Ashkenazi are scoring 26 percent ashkenazi on that, there's some weird sorcery behind that thing we may never figure out. Normal calculators with an ashkenazi component give 100 percent or perhaps a bit less of that to ashkenazi users.

If it's giving real Ashkenazi Jews 26%, then it's crap and should be taken off gedmatch.

 

[Blanco]

Eurogenes calculators are one of the best because of the high SNP numbers.160-180k SNP is good for a Gedmatch calculator. Harappaworld is also good and better for Mediterranean and Near Eastern populations because Eurogenes was specified for people of Predominately European ancestry.

PuntDNAL works well for the Calabrese because it's made for Mediterranean populations: It's great to see how they remained pure from Indo-European, Slavic invasions unlike the Northern parts of Mainland Greece that's why they get closer to even Cyprus than most parts of Mainland Greece even. East Med brotherhood:grin:

#	Population	Percent
1	SW_Europe	37.96
2	West_Asia	27.15
3	SW_Asia	18.14
4	NE_Europe	13.4
5	SE_Asia	2.19
6	East_Africa	0.92
7	Oceania	0.17
8	South_Africa	0.07

Single Population Sharing:

#	Population (source)	Distance
1	Sephardic_Jew	3.88
2	Italian_Sicilian	4.49
3	Ashkenazy_Jew	4.72
4	Italian_Abruzzo	9.55
5	Cypriot	9.63
6	Greek_Central	9.72
7	Turkish	13.08
8	Lebanese_Christian	14.82
9	Turkish_Kayseri	14.94
10	Syrian	15.02
11	Greek_Thessaly	15.07
12	Lebanese_Druze	15.45
13	Albanian	15.76
14	Lebanese_Muslim	16.41
15	Turkish_Aydin	16.87
16	Kosovar	16.92
17	Jordanian	16.95
18	Italian_Tuscan	17.14
19	Palestinian	19.72
20	Armenian	20.08

Mixed Mode Population Sharing:

#		Primary Population (source)	Secondary Population (source)	Distance
1		69.4%	Ashkenazy_Jew	+	30.6%	Cypriot	@	2.29
2		71.3%	Italian_Sicilian	+	28.7%	Cypriot	@	2.45
3		79.6%	Italian_Sicilian	+	20.4%	Lebanese_Christian	@	2.49
4		83.9%	Ashkenazy_Jew	+	16.1%	Samaritan_Jew	@	2.56
5		80.6%	Italian_Sicilian	+	19.4%	Lebanese_Druze	@	2.58
6		78.7%	Ashkenazy_Jew	+	21.3%	Lebanese_Christian	@	2.6
7		85%	Italian_Sicilian	+	15%	Samaritan_Jew	@	2.62
8		81.8%	Italian_Sicilian	+	18.2%	Lebanese_Muslim	@	2.66
9		85%	Italian_Sicilian	+	15%	Assyrian	@	2.74
10		79.9%	Ashkenazy_Jew	+	20.1%	Lebanese_Druze	@	2.77
11		50.2%	Italian_Abruzzo	+	49.8%	Cypriot	@	2.78
12		80.8%	Italian_Sicilian	+	19.2%	Syrian	@	2.79
13		81.1%	Ashkenazy_Jew	+	18.9%	Lebanese_Muslim	@	2.85
14		71.6%	Lebanese_Christian	+	28.4%	French_Basque	@	2.86
15		85.2%	Italian_Sicilian	+	14.8%	Armenian	@	2.87
16		70.7%	Lebanese_Druze	+	29.3%	French_Basque	@	2.87
17		64.8%	Cypriot	+	35.2%	Italian_Tuscan	@	2.89
18		75.6%	Cypriot	+	24.4%	Spaniard	@	2.9
19		79.9%	Ashkenazy_Jew	+	20.1%	Syrian	@	2.92
20		89.4%	Sephardic_Jew	+	10.6%	Balkar	@	2.99

 

[Angela]

If your algorithm is lousy, and your clusters are wrong it doesn't matter how many snps you use. The proof is in the pudding as we say. His calculators can't recognize a Tuscan accurately, or Jews, so they're crap for at least Italians and Jews.

Is this a joke? A calculator is good for Calabrians which doesn't have a southern Italian reference sample? Are you aware that if there's no Southern Italian or Calabrian reference sample, it can't come up in the prediction? Therefore we can't know if it captures Calabrians adequately or not? Get it?

I feel like Alice after she fell down the rabbit hole.

Could you telegraph that you're a t-roll any more clearly? You know I'm half quasi-Tuscan, and my husband is half Calabrian and so you present calculators that show we're both high in "Jewish" ancestry?

So what? Is that supposed to hurt my feelings? (That would be absolutely fine with both me and my husband if it were true, btw.) Go away and find something productive to do with your time other than t-rolling Italians you idiotic racist.

Keep going and you'll start getting infractions for posting provocative, disruptive material.

 

[Blanco]

If your algorithm is lousy, and your clusters are wrong it doesn't matter how many snps you use. The proof is in the pudding as we say. His calculators can't recognize a Tuscan accurately, or Jews, so they're crap for at least Italians and Jews.

It's true. However most users who tested Eurogenes K13-K15 got accurate results for the most part. Altough Eurogenes calulators were created for people of North-Central European ancestry, therefore it will be less accurate for Mediterranean populations. Still Eurogenes is better than Gedrosia, MDLP calculators. I do agree with you that Polako is often biased and Nordicist in many terms, given his Skadi past. I don't symphatize with him, but neither with the Nordicist ideas which try to throw any connections between Italians and Eastern Mediterranean population under the bus.

Could you telegraph that you're a t-roll any more clearly? You know I'm half quasi-Tuscan, and my husband is half Calabrian and so you present calculators that show we're both high in "Jewish" ancestry?

I don't know you nor your heritage, and I don't know why it has anything to do with this topics. I just posted these results because I'm interested of it.:cool-v: I never claimed that Italians have Jewish ancestry I just said once in my post that due to shared Neolithic Mediterranean ancestry these populations can get similar results. There are no genetic Jews, but due to close intermarriages and with this being said most Ashkenazi Jews are closely related given by the bottleneck effect they're genetically a district, yet similar populations to those who carry high Anatolian Neolithic, Armenian EBA, Barchin Neolithic and other ancestral links high in Basal Eurasian with moderately to high Villaburra related ancestry.

So what? Is that supposed to hurt my feelings? (That would be absolutely fine with both me and my husband if it were true, btw.) Go away and find something productive to do with your time other than t-rolling Italians you idiotic racist.

Italians are great people and in fact I think they're the most admired and envied people in the Western heisphere. I have friends from Italy and they're absolutely chill and good hard working people. Am I racist? how I can't be racist against my own akin, I just don't support the Eurocentric point of view which considers the Mesolithic-Paleothic rich Indo European populations more desirable than the Basal-Neolithic rich populations who brought farming and created the European civillizations. By the way the term Europe is also from the Near East via by Phoenicians.

 

[AdeoF]

K15 is not that accurate it says that my first is Portuguese from the Single Population Sharing. Yeah by looking at it it's not very good it needs to be removed also it's not accurate for both northern Italians and Jews.

 

[Pratt]

In the PCA the Tuscan kit goes close to the Spaniards, while Catanzaro is south of Southern Italian samples. Unusual results or bad calculator?

 

[Angela]

Pratt;535908]In the PCA the Tuscan kit goes close to the Spaniards, while Catanzaro is south of Southern Italian samples. Unusual results or bad calculator?

What do you mean by unusual results? Do you mean unrepresentative samples?

Is the Catanzaro sample the one that Lazaridis said behaved strangely and might be an extremely drifted?

The HGDP samples are the ones chosen by Cavalli-Sforza aren't they?

Something is definitely wrong, however.

[/QUOTE]

 

[davef]

quadrant 1 is Celtic Germanic land, q2 is Iberian Sardinian north Italy land, q3 is East Med land, q4 is Slavic land

q5 is candy land

 

[Blanco]

From Nicotera, Calabria: T227284

#	Population	Percent
1	East_Med	30.19
2	West_Asian	17.14
3	West_Med	16.4
4	Atlantic	15.66
5	Red_Sea	6.75
6	North_Sea	6.56
7	Baltic	3.78
8	Northeast_African	1.49
9	Eastern_Euro	1.44
10	South_Asian	0.59

Single Population Sharing:

#	Population (source)	Distance
1	South_Italian	4.1
2	East_Sicilian	5.58
3	Sephardic_Jewish	5.69
4	Central_Greek	6.33
5	Italian_Jewish	6.67
6	Italian_Abruzzo	7.25
7	West_Sicilian	7.71
8	Algerian_Jewish	8.13
9	Ashkenazi	8.32
10	Tunisian_Jewish	9.91
11	Greek	11.71
12	Cyprian	12.14
13	Tuscan	13.41
14	Libyan_Jewish	13.5
15	Greek_Thessaly	13.52
16	Lebanese_Muslim	15.09
17	Syrian	16.32
18	Turkish	16.36
19	Samaritan	18.89
20	Bulgarian	19.51

Mixed Mode Population Sharing:

#		Primary Population (source)	Secondary Population (source)	Distance
1		70.5%	Cyprian	+	29.5%	Spanish_Castilla_La_Mancha	@	2.89
2		63.7%	Lebanese_Muslim	+	36.3%	Spanish_Andalucia	@	2.91
3		75.4%	West_Sicilian	+	24.6%	Georgian_Jewish	@	2.91
4		76.9%	Cyprian	+	23.1%	French_Basque	@	2.99
5		74.2%	West_Sicilian	+	25.8%	Assyrian	@	3
6		64.2%	Italian_Abruzzo	+	35.8%	Cyprian	@	3.11
7		72.7%	Lebanese_Muslim	+	27.3%	French_Basque	@	3.14
8		68.8%	Cyprian	+	31.2%	Spanish_Andalucia	@	3.15
9		72%	Cyprian	+	28%	Spanish_Aragon	@	3.16
10		64.8%	South_Italian	+	35.2%	Sephardic_Jewish	@	3.2
11		54.3%	Kurdish_Jewish	+	45.7%	Spanish_Andalucia	@	3.21
12		58.1%	Sephardic_Jewish	+	41.9%	Italian_Abruzzo	@	3.21
13		67.1%	Lebanese_Muslim	+	32.9%	Spanish_Aragon	@	3.22
14		70.2%	Cyprian	+	29.8%	Spanish_Valencia	@	3.22
15		76.1%	West_Sicilian	+	23.9%	Armenian	@	3.22
16		70.4%	Cyprian	+	29.6%	Spanish_Castilla_Y_Leon	@	3.24
17		59.2%	Italian_Abruzzo	+	40.8%	Tunisian_Jewish	@	3.31
18		65.2%	Lebanese_Muslim	+	34.8%	Spanish_Valencia	@	3.32
19		75.8%	West_Sicilian	+	24.2%	Kurdish_Jewish	@	3.35
20		72.5%	Cyprian	+	27.5%	Spanish_Cantabria	@	3.37

From Carfizzi, Calabria

#	Population	Percent
1	East_Med	29.99
2	West_Med	18.27
3	Atlantic	12.61
4	North_Sea	11.65
5	West_Asian	9.27
6	Baltic	6.3
7	Red_Sea	4.98
8	Eastern_Euro	4.97
9	Northeast_African	1.15
10	Siberian	0.81

Single Population Sharing:

#	Population (source)	Distance
1	East_Sicilian	6.27
2	West_Sicilian	6.46
3	Ashkenazi	6.65
4	Central_Greek	6.7
5	South_Italian	7.08
6	Italian_Jewish	7.85
7	Italian_Abruzzo	8.06
8	Greek_Thessaly	8.87
9	Greek	9.15
10	Algerian_Jewish	9.18
11	Sephardic_Jewish	10.01
12	Tuscan	10.65
13	Tunisian_Jewish	13.46
14	Libyan_Jewish	13.61
15	Bulgarian	16.28
16	Cyprian	16.46
17	North_Italian	16.49
18	Romanian	18.18
19	Lebanese_Muslim	20.22
20	Serbian	20.61

Mixed Mode Population Sharing:

#		Primary Population (source)	Secondary Population (source)	Distance
1		83.3%	Italian_Jewish	+	16.7%	Ukrainian	@	3.69
2		51%	Greek_Thessaly	+	49%	Algerian_Jewish	@	3.75
3		53.8%	Italian_Jewish	+	46.2%	Greek_Thessaly	@	3.75
4		80.1%	Italian_Jewish	+	19.9%	Hungarian	@	3.82
5		74.8%	Italian_Jewish	+	25.2%	Serbian	@	3.86
6		79.3%	Italian_Jewish	+	20.7%	Moldavian	@	3.89
7		83.2%	Italian_Jewish	+	16.8%	Ukrainian_Lviv	@	3.9
8		85.9%	Italian_Jewish	+	14.1%	Estonian	@	3.91
9		83.5%	Italian_Jewish	+	16.5%	South_Polish	@	3.91
10		76.9%	Algerian_Jewish	+	23.1%	Hungarian	@	3.94
11		80.6%	Italian_Jewish	+	19.4%	Croatian	@	3.94
12		84.5%	Italian_Jewish	+	15.5%	Polish	@	3.95
13		85.3%	Italian_Jewish	+	14.7%	East_Finnish	@	4.02
14		71%	Algerian_Jewish	+	29%	Serbian	@	4.04
15		85.1%	Italian_Jewish	+	14.9%	Finnish	@	4.05
16		85.4%	Italian_Jewish	+	14.6%	Kargopol_Russian	@	4.1
17		85%	Italian_Jewish	+	15%	Southwest_Finnish	@	4.11
18		82.3%	Italian_Jewish	+	17.7%	East_German	@	4.12
19		85.5%	Italian_Jewish	+	14.5%	Estonian_Polish	@	4.13
20		82.5%	Algerian_Jewish	+	17.5%	North_Swedish	@	4.15

Both results on Eurogenes K15 PCA plot:

 

[Angela]

Is this why my husband is so smart? You people traffic in stereotypes, so has that occurred to you? I'm sure you'd say that's why he's so successful in business too, yes?

My God, people, get a life.

 

[Pratt]

What do you mean by unusual results? Do you mean unrepresentative samples?

Is the Catanzaro sample the one that Lazaridis said behaved strangely and might be an extremely drifted?

The HGDP samples are the ones chosen by Cavalli-Sforza aren't they?

Something is definitely wrong, however.

I meant that it shouldn't be there in a PCA, in that position. The PCA is better than the Oracle, the latter is more deceptive. The Oracle for the individual Tuscan HGDP chooses all Spaniards as the first population (and the Spaniards in a PCA are in a secluded position, certainly not at the center of a PCA) and even the Oracle chooses a super-isolated population like the Basques, so then the Oracle is forced to choose a second population much more to the south-east. Oracle in this way demonstrates all its frailty by failing to choose a more adequate population. And perhaps this is also due to the spreadsheet of the K15 not specifically to the software or to weird results. From the looks of it the individual Tuscan HGDP looks like a Sardinian-shifted North Italian, or even a Sardinian-shifted Tuscan.

Catanzaro is not an academic sample, according to Gedmatch she is a person from Decollatura, Catanzaro, Calabria, while HGDP are Stanford University's samples as I know.

To understand something more I did another Eurogenes K15 PCA with all the HGDP academic samples in it, both Bergamo and Tuscan HGDP, including the one posted above. And with some surprise, not only 5 individuals Tuscan HGDP have a shift towards Sardinia, but also 5 individuals of Bergamo HGDP have this shift, with all very high atlantic and north sea that is instead lowered, exactly like HGDP posted by OP, so it's a common problem that both these samples have. While the rest of HGDP samples plot where are supposed to be, with a more balanced ratio between north sea and atlantic.

I did a further test. I added to the PCA 10 people from Lombardy and 10 people from Tuscany, non academic samples, and none of these individuals shifts in this way and they all seem to plot correctly.

You're right. There is something definitely wrong. It could be something wrong with raw data of some of the HGDP samples. I suppose something went wrong in the process of converting from the academic format to a format suitable for uploading to gedmatch, probably low-coverage data. Or who knows, maybe it's something wrong with this calculator. We should see how the same samples behave in the other calculators.

 

[Angela]

I don't get why 5 Tuscan and 5 Bergamo academic samples would behave so strangely on a PCA. If something went wrong during the transfer to gedmatch wouldn't all the samples have been effected? Why only a portion of them?

I wonder if some of the samples uploaded to gedmatch are the actual HGDP samples? Is there any way to check?

As you say, it's either that, or the calculator is even more flawed than I suspected, or both.

As to Oracle, it's not that it's wrong, imo, it's that people don't know how to interpret it, or they choose to interpret it incorrectly for their own reasons. As you say, the algorithm tells you the "direction" perhaps, of where the gene flow is coming from. It's not saying that exact population was part of an admixture event that resulted in the subject population.

I was explaining this to people on 23andme six years ago, the same people, some of them, btw, some in the context of English results where they'd get a French percentage, for example, and think that meant actual French admixture in their family tree. . Either they're just remarkably stupid, or they just pretend not to understand.

 

[Pratt]

I don't get why 5 Tuscan and 5 Bergamo academic samples would behave so strangely on a PCA. If something went wrong during the transfer to gedmatch wouldn't all the samples have been effected? Why only a portion of them?

I wonder if some of the samples uploaded to gedmatch are the actual HGDP samples? Is there any way to check?

As you say, it's either that, or the calculator is even more flawed than I suspected, or both.

The academic samples, especially in the past, are never released in a format compatible with Gedmatch, they need to be converted to a build 37 format. I think something may have gone wrong in this conversion process.

 

[Hauteville]

Carfizzi is Arbereshe

 

[Angela]

The academic samples, especially in the past, are never released in a format compatible with Gedmatch, they need to be converted to a build 37 format. I think something may have gone wrong in this conversion process.

I still find it a bit odd that only some of the samples were effected. Why not all?

Also, if this could happen with some of the Italian samples, it could happen with samples from other countries.

One more reason for me not to place too much confidence in gedmatch results.

 

[davef]

Why didn't these conversion issues show up in other Southern European populations like the Greeks and Iberians who have a lot in common with Italians? Just curious

 

[Pratt]

I still find it a bit odd that only some of the samples were effected. Why not all?

Also, if this could happen with some of the Italian samples, it could happen with samples from other countries.

One more reason for me not to place too much confidence in gedmatch results.

Of course it's odd.

Why didn't these conversion issues show up in other Southern European populations like the Greeks and Iberians who have a lot in common with Italians? Just curious

There are no Greeks and Iberians on HGDP to begin with.